How to diagnose a (potentially) defective T42p fan?

[ginahoy]

After the original CPU fan started making noise, I installed a new fan to the existing heat sink (T42p with short heat sink). The fan failed after just a few weeks. Upon further inspection, I discovered the fan fuse (F4) had failed.

It seems reasonable to assume this was no coincidence and that the fan caused a the fuse to blow, right?

Upon close inspection, the fan leads were not pinched and the insulation is unblemished. So I tested the leads expecting to find an internal short. Surprisingly, no short. The resistances are as follows:

10k ohms between outer two leads (red & brown on this fan)
10k ohms between center lead (blue) and brown outer lead
375 ohms between center lead and red outer lead

If anyone has a spare fan on hand that's known to be good, I would appreciate it if you could ohm it out for me. Why do these fans have 3 leads? If I want to test the fan with a battery, to which leads do I apply power? Presumably the red lead is positive.

Thanks

[rkawakami]

Not sure what the convention is for identifying power and ground leads in this case, but I believe that the "third" wire is for detecting the fan speed. Typically a hall-effect sensor is used to determine the position of the fan as it spins.

If I were to guess, I'd say that the 375 ohms is the coil winding of the fan's motor, with the red wire the +12V and the blue ground. That leaves the brown the sensor. If you measure (ohms) the brown and blue wires and are able to turn the fan, you may see "pulses" on the meter.

edit: Sorry, maybe the fan is +5V instead... I'll have to check some docs later on.

[ajkula66]

(T42p with short heat sink).

No T4*p system calls for a short heatsink. Long only.

[ginahoy]

No T4*p system calls for a short heatsink. Long only.

My bad, it's a T42. I forgot the monitor is not original. It came from a T42P system.

I believe that the "third" wire is for detecting the fan speed...

Right. I use SpeedFan to manage several case fans in my tower desktop. Those fans have 3 wires. That must be the industry standard, even for fans that aren't used on monitored circuits.

If I were to guess, I'd say that the 375 ohms is the coil winding of the fan's motor, with the red wire the +12V and the blue ground. That leaves the brown the sensor.

Ok, now the readings make sense. I just powered up the fan with a pair of 3V button cells and the fan looks like it's OK.

So the question is, what could have blown the fan fuse? Anyone care to speculate? I don't buy coincidence: suddenly 3 wks after changing fan, the fuse blows?

The one thing I can't verify is if the replacement fan moves as much air as the OEM fan. However, a hot CPU wouldn't cause an overcurrent situation in the fan circuit, right?

[RealBlackStuff]

If you replaced the fan and put in no, or not enough, or too much, fresh thermal paste, the poor fan would be constantly working at top speed!
That would blow anybody's circuit!

[ginahoy]

If you replaced the fan and put in no, or not enough, or too much, fresh thermal paste, the poor fan would be constantly working at top speed!
That would blow anybody's circuit!

Not sure if you're trying to be funny but that's not how electronics work. First, the F4 fuse is rated at 500 milliamps. There's no way those little fans could possibly exceed that unless there's a dead short. Depending on whether the circuit is 12 or 5 volts, that fan will draw no more than 32 milliamps (I = V / R). The only purpose of the fuse is to protect upstream components from a short or externally applied current. Also, a fan's current doesn't increase with run time so if the CPU was overheating, the fan would simply operate at it's maximum speed and current, which is more than an order of magnitude less than the fuse rating.

The processor may have been overheating if the fan was an underperformer, but not because of the thermal paste. I use Arctic Silver and fully understand the application procedure.

Excess heat *could* affect a fuse, but that seems unlikely. The sm fuses used on these boards have a high melting point relative to their current rating. I would think other components would be much more affected by heat.

The purpose of this thread was to understand how to test the fan. That's done. Beyond that, my curiosity as an engineer wants to know why the fuse failed, but in the end it doesn't really matter.

Thanks guys for the assist.

[shawross]

My 2 cents worth is I would be suspicious of that fan and or wiring plain and simple.

Goodluck

[rkawakami]

Fan is 5V; pin 1 on the connector or the red wire. Ground is pin 2 (middle) and speed sensor is pin 3.

As to why the F4 fuse blew out, it could be due to some of the other parts in the circuit. In addition to the fan, F4 protects a power MOSFET (Q1 - FDC658P) which is used to supply the 5V to the fan. Q1 is gated by Q77, a DTC114EE transistor that is basically acting like an inverter so that when the input goes low, it delivers a high to the MOSFET, turning it on and that allows the 5V to get to the fan. As far as I can tell, the maximum current that could occur if Q77 shorts out, would be something like 4.5mA. This assumes that the 1K and 100 ohm resistors in the circuit are okay. If your measurement of the fan says 375 ohms, then the fan will draw about 13mA. Like you suggest, not enough current to blow the fuse.

A component that is connected directly across the +5V and system ground, besides the fan itself, is a capacitor (C301). If for some reason it has an internal short, then the 5V gets dumped through F4 and Q1 into ground. The specs for Q1 says it can handle up to 4 amps so that means the fuse should blow before the MOSFET. Of course any other short on that power line to ground could also cause the fuse to go ppft. It may be that the cap shorted out and in the process, fully opened up (i.e., have infinite resistance as a capacitor normally should have but it's also no longer a capacitor). I doubt that the fan could have a temporary short (less than 10 ohms to get the 0.5A needed to kill the fuse) and still continue to work so I'd suspect the cap or a momentary short somewhere else in the fan circuit.

[ginahoy]

Thanks for that analysis rkawakami. It makes sens that something upstream of the fuse could also take it out. I guess whatever fault took out the fuse could have been caused by CPU overheating due to under-performing (read cheap) fan. But whatever it was, it's either transient or non-essential because I when I got the CPU fan error message on POST, I pressed escape and it proceeded to boot into Windows. I shut it down to avoid damaging the CPU.

I have since bought Neil's T42P. His has the 745 so I plan to swap out with my 765. If it was damaged, I'll know soon enough. Still, I'd like to monitor CPU temperature. Is there a utility for that?

[RealBlackStuff]

https://sourceforge.net/projects/tp4xfancontrol/

[rkawakami]

I use a simple one on most of my older systems but it has been abandoned for several years: MobileMeter

You can find it at Softpedia: http://www.softpedia.com/get/System/Sys ... eter.shtml but there have been some problems reported in the past trying to download from there.

The other place I know of is here: http://www.geocities.co.jp/SiliconValle ... mm0310.zip I just downloaded the file from both places and compared them to the file I obtained in 2009 and they are the same; MD5 checksum is ea9f28949e96d31977f95d1f6af84faa.

It was written to run under Windows 2000 and XP. I don't think any of my Windows 7 systems have it so I don't know if it will run there. Besides monitoring CPU temperature it can also watch the clock frequency, the status of the battery/charging system and the temperature of the hard drive(s).

[ginahoy]

Thanks for the referrals.

[rkawakami]

Here's an example of what MobileMeter can look like:

Transparent Operation and Always On Top modes are enabled. This is running on an A31p. Top line is clock speed, 2nd line is CPU temp, 3rd is charging system, 4th is HDD temp. I've set it to just display the values. There's an option to get graphical plots so you can see how the values are trending. I will have to say on very rare occasions I have seen the program eat up CPU cycles and I've had to kill and restart it. Maybe once a year but the system was also acting squirrelly anyway. I can't say if it was the cause or an effect. YMMV.

[ginahoy]

I run a similar tsr on my desktop - speedfan - but I don't think it's compatible with Thinkpads. I custom built my desktop system specifically to be as silent as possible. It's 2006 technology (AMD Opteron 185 2.6 Ghz dual core), and still my main workhorse computer. I installed super quiet fans on the CPU cooler block, rear case, front lower case (for HDD's) and bottom of case (for fanless graphics card). I keep fans at a minimum speed 90% of the time. They automatically speed up when a specific component temperature exceeds the threshold I set. The one fan I don't control is the one that's part of the power supply, as I didn't see a way to monitor and control it. Fun stuff.